Graph Neural Networks (GNNs) with equivariant properties have achieved significant success in modeling complex dynamic systems and molecular properties. However, their expressiveness ability is limited by: (1) Existing methods often overlook the over-smoothing issue caused by traditional GNN models, as well as the gradient explosion or vanishing problems in deep GNNs. (2) Most models operate on first-order information, neglecting that the real world often consists of second-order systems, which further limits the model’s representation capabilities. To address these issues, we propose the Dual Second-order Equivariant Graph Ordinary (DuSEGO) Differential Equation for equivariant representation. Specifically, DuSEGO applies the dual second-order equivariant graph ordinary differential equations (Graph ODEs) to both graph embeddings and node coordinates simultaneously. Theoretically, we first prove that DuSEGO maintains the equivariant property. Furthermore, we provide theoretical insights showing that DuSEGO effectively alleviates the over-smoothing problem in both feature representation and coordinate update. Additionally, we demonstrate that the proposed DuSEGO mitigates the exploding and vanishing gradients problem, facilitating the training of deep multi-layer GNNs. Extensive experiments on benchmark datasets validate the superiority of the proposed DuSEGO compared to baselines.

Recent LLMs exhibit limited effectiveness on molecular property prediction task due to the semantic gap between molecular representations and natural language, as well as the lack of domain-specific knowledge.To address these challenges, we propose MolRAG, a Retrieval-Augmented Generation framework integrating Chain-of-Thought reasoning for molecular property prediction.MolRAG operates by retrieving structurally analogous molecules as contextual references to guide stepwise knowledge reasoning through chemical structureproperty relationships.This dual mechanism synergizes molecular similarity analysis with structured inference, while generating humaninterpretable rationales grounded in domain knowledge.Experimental results show Mol-RAG outperforms pre-trained LLMs on four datasets, and even matches supervised methods, achieving performance gains of 1.1%-45.7%over direct prediction approaches, demonstrating versatile effectiveness.Our code is available at https://github.com/AcaciaSin/MolRAG.

Approximation-based spectral graph neural networks, which construct graph filters with function approximation, have shown substantial performance in graph learning tasks. Despite their great success, existing works primarily employ polynomial approximation to construct the filters, whereas another superior option, namely ration approximation, remains underexplored. Although a handful of prior works have attempted to deploy the rational approximation, their implementations often involve intensive computational demands or still resort to polynomial approximations, hindering full potential of the rational graph filters. To address the issues, this paper introduces ERGNN, a novel spectral GNN with explicitly-optimized rational filter. ERGNN adopts a unique two-step framework that sequentially applies the numerator filter and the denominator filter to the input signals, thus streamlining the model paradigm while enabling explicit optimization of both numerator and denominator of the rational filter. Extensive experiments validate the superiority of ERGNN over state-of-the-art methods, establishing it as a practical solution for deploying rational-based GNNs.

Existing preference optimization objectives for language model alignment require additional hyperparameters that must be extensively tuned to achieve optimal performance, increasing both the complexity and time required for fine-tuning large language models. In this paper, we propose a simple yet effective hyperparameter-free preference optimization algorithm for alignment. We observe that promising performance can be achieved simply by optimizing inverse perplexity, which is calculated as the inverse of the exponentiated average log-likelihood of the chosen and rejected responses in the preference dataset. The resulting simple learning objective, SimPER, is easy to implement and eliminates the need for expensive hyperparameter tuning and a reference model, making it both computationally and memory efficient. Extensive experiments on widely used real-world benchmarks, including MT-Bench, AlpacaEval 2, and 10 key benchmarks of the Open LLM Leaderboard with 5 base models, demonstrate that SimPER consistently and significantly outperforms existing approaches—even without any hyperparameters or a reference model. For example, despite its simplicity, SimPER outperforms state-of-the-art methods by up to 5.7 points on AlpacaEval 2 and achieves the highest average ranking across 10 benchmarks on the Open LLM Leaderboard. The source code for SimPER is publicly available at: https://github.com/tengxiao1/SimPER.

Muhammad Arslan Manzoor, Ruihong Zeng, Dilshod Azizov, Preslav Nakov, Shangsong Liang. Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers). 2025.

Spectral graph neural networks are proposed to harness spectral information inherent in graph-structured data through the application of polynomial-defined graph filters, recently achieving notable success in graph-based web applications. Existing studies reveal that various polynomial choices greatly impact spectral GNN performance, underscoring the importance of polynomial selection. However, this selection process remains a critical and unresolved challenge. Although prior work suggests a connection between the approximation capabilities of polynomials and the efficacy of spectral GNNs, there is a lack of theoretical insights into this relationship, rendering polynomial selection a largely heuristic process.

Recently, because of the high-quality representations of contrastive learning methods, rehearsal-based contrastive continual learning has been proposed to explore how to continually learn transferable representation embeddings to avoid the catastrophic forgetting issue in traditional continual settings. Based on this framework, we propose Contrastive Continual Learning via Importance Sampling (CCLIS) to preserve knowledge by recovering previous data distributions with a new strategy for Replay Buffer Selection (RBS), which minimize estimated variance to save hard negative samples for representation learning with high quality. Furthermore, we present the Prototype-instance Relation Distillation (PRD) loss, a technique designed to maintain the relationship between prototypes and sample representations using a self-distillation process. Experiments on standard continual learning benchmarks reveal that our method notably outperforms existing baselines in terms of knowledge preservation and thereby effectively counteracts catastrophic forgetting in online contexts. The code is available at https://github.com/lijy373/CCLIS.

The simultaneous use of multiple drugs may enhance the drug therapy process but can also increase the risk of adverse side effects. It is critical to identify the drug-drug interactions (DDIs) to avoid the possible side effects. Deep learning-based DDIs prediction models heavily rely on the quantity and quality of training data and labeled DDIs require extensive experiments, demanding significant time and resources. With the rapid discovery of new drugs, it is necessary to design a label-efficient learning method to mitigate the challenge of high DDIs annotation costs. In this study, we seamlessly integrate active learning into the multi-class DDIs prediction and present Margin-based Dynamic Cluster (MDC) active learning strategy. Extensive experiments show that our method significantly reduces annotation costs while maintaining model performance and outperforms other approaches.

Pre-trained models such as BERT have achieved remarkable results in text matching tasks. However, existing models still suffer from the challenge of capturing local subtle differences when modeling complex semantic matching relationships. In this work, we find that the integration of local syntax awareness and global semantics is crucial for text matching. Meanwhile, we propose the Local and Global Syntax Graph Calibration (LG-SGC) module, which can explore local syntactic and global semantic information for the matching task. Specifically, we first introduce an auxiliary task inside BERT to capture local subtle grammatical differences. Then, we retain the original attention operation to capture global matching features. Finally, we design an information fusion layer to effectively combine local and global information to deepen the understanding of the matching task. We conduct extensive experiments on 10 benchmarks, and LG-SGC significantly outperforms previous models.

Inverse molecular generation is an essential task for drug discovery, and generative models offer a very promising avenue, especially when diffusion models are used. Despite their great success, existing methods are inherently limited by the lack of a semantic latent space that can not be navigated and perform targeted exploration to generate molecules with desired properties. Here, we present a property-guided diffusion model for generating desired molecules, which incorporates a sophisticated diffusion process capturing intricate interactions of nodes and edges within molecular graphs and leverages a time-dependent molecular property classifier to integrate desired properties into the diffusion sampling process. Furthermore, we extend our model to a multi-property-guided paradigm. Experimental results underscore the competitiveness of our approach in molecular generation, highlighting its superiority in generating desired molecules without the need for additional optimization steps.

Transformer-based Large Language Models (LLMs) are pioneering advances in many natural language processing tasks, however, their exceptional capabilities are restricted within the preset context window of Transformer. Position Embedding (PE) scaling methods, while effective in extending the context window to a specific length, demonstrate either notable limitations in their extrapolation abilities or sacrificing partial performance within the context window. Length extrapolation methods, although theoretically capable of extending the context window beyond the training sequence length, often underperform in practical long-context applications. To address these challenges, we propose Continuous Length EXtrapolation (CLEX) for LLMs. We generalise the PE scaling approaches to model the continuous dynamics by ordinary differential equations over the length scaling factor, thereby overcoming the constraints of current PE scaling methods designed for specific lengths. Moreover, by extending the dynamics to desired context lengths beyond the training sequence length, CLEX facilitates the length extrapolation with impressive performance in practical tasks. We demonstrate that CLEX can be seamlessly incorporated into LLMs equipped with Rotary Position Embedding, such as LLaMA and GPT-NeoX, with negligible impact on training and inference latency. Experimental results reveal that CLEX can effectively extend the context window to over 4× or almost 8× training length, with no deterioration in performance. Furthermore, when evaluated on the practical LongBench benchmark, our model trained on a 4k length exhibits competitive performance against state-of-the-art open-source models trained on context lengths up to 32k. Our code is available at https://github.com/DAMO-NLP-SG/CLEX.

Recent methods are proposed to improve performance of domain adaptation by inferring domain index under an adversarial variational bayesian framework, where domain index is unavailable. However, existing methods typically assume that the global domain indices are sampled from a vanilla gaussian prior, overlooking the inherent structures among different domains. To address this challenge, we propose a Bayesian Domain Adaptation with Gaussian Mixture Domain-Indexing(GMDI) algorithm. GMDI employs a Gaussian Mixture Model for domain indices, with the number of component distributions in the "domain-themes'' space adaptively determined by a Chinese Restaurant Process. By dynamically adjusting the mixtures at the domain indices level, GMDI significantly improves domain adaptation performance. Our theoretical analysis demonstrates that GMDI achieves a more stringent evidence lower bound, closer to the log-likelihood. For classification, GMDI outperforms all approaches, and surpasses the state-of-the-art method, VDI, by up to 3.4%, reaching 99.3%. For regression, GMDI reduces MSE by up to 21% (from 3.160 to 2.493), achieving the lowest errors among all methods.

Understanding the genetic basis of disease is a fundamental aspect of medical research, as genes are the classic units of heredity and play a crucial role in biological function. Identifying associations between genes and diseases is critical for diagnosis, prevention, prognosis, and drug development. Genes that encode proteins with similar sequences are often implicated in related diseases, as proteins causing identical or similar diseases tend to show limited variation in their sequences. Predicting gene-disease association (GDA) requires time-consuming and expensive experiments on a large number of potential candidate genes. Although methods have been proposed to predict associations between genes and diseases using traditional machine learning algorithms and graph neural networks, these approaches struggle to capture the deep semantic information within the genes and diseases and are dependent on training data. To alleviate this issue, we propose a novel GDA prediction model named FusionGDA, which utilizes a pre-training phase with a fusion module to enrich the gene and disease semantic representations encoded by pre-trained language models. Multi-modal representations are generated by the fusion module, which includes rich semantic information about two heterogeneous biomedical entities: protein sequences and disease descriptions. Subsequently, the pooling aggregation strategy is adopted to compress the dimensions of the multi-modal representation. In addition, FusionGDA employs a pre-training phase leveraging a contrastive learning loss to extract potential gene and disease features by training on a large public GDA dataset. To rigorously evaluate the effectiveness of the FusionGDA model, we conduct comprehensive experiments on five datasets and compare our proposed model with five competitive baseline models on the DisGeNet-Eval dataset. Notably, our case study further demonstrates the ability of FusionGDA to discover hidden associations effectively. The complete code and datasets of our experiments are available at https://github.com/ZhaohanM/FusionGDA.

Abstract Vision Transformers for pre-trained models explore semantic context and spatial relationships for images, which heavily depend on how you select image patches. In this paper, we propose a novel Spatial-aware Multi-directional Patches Multi-cycle Autoencoder (SMPMA) for a pre-trained model that brings the following benefits: (1) Spatial-aware Multi-directional (SM) patches are created with multi-directional spatial locations, transforming a whole image autoencoder problem into a short-span image patches autoencoder problem; (2) SM patches admit a self-cycle autoencoder alignment learning for the first stage and a cross-cycle interaction learning for the second stage, which makes patches align and interact for optimization; (3) SM patches enable to explore local object features and correlation distribution of adjacent pixels by taking arbitrary sampled patches as inputs. Experimental results on four downstream tasks show that our model can achieve state-of-the-art performance over the tasks of image generation, image classification and semantic segmentation.

Due to the easy access, implicit feedback is often used for recommender systems. Compared with point-wise learning and pair-wise learning methods, list-wise rank learning methods have superior performance for top- \(N\) recommendation. Recent solutions, especially the list-wise methods, simply treat all interacted items of a user as equally important positives and annotate all no-interaction items of a user as negatives. For the list-wise approaches, we argue that this annotation scheme of implicit feedback is over-simplified due to the sparsity and missing fine-grained labels of the feedback data. To overcome this issue, we revisit the so-called positive and negative samples. First, considering the loss function of list-wise ranking, we analyze the impact of false positives and negatives theoretically. Second, based on the observation, we propose a self-adjusting credibility weight mechanism to re-weigh the positive samples and exploit the higher-order relation based on item–item matrix to sample the critical negative samples. In order to prevent the introduction of noise, we design a pruning strategy for critical negatives. Besides, to combine the reconstruction loss function for the positive samples and critical negative samples, we develop a simple yet effective VAEs framework with linear structure, which abandons the complex non-linear structure. Extensive experiments are conducted on six public real-world datasets. The results demonstrate that, our VAE* outperforms other VAE-based models by a large margin. Besides, we also verify the effect of denoising positives and exploring critical negatives by ablation study.

In this article, we study the problem of embedding temporal attributed networks, with the goal of which is to learn dynamic low-dimensional representations over time for temporal attributed networks. Existing temporal network embedding methods only learn the representations for nodes, which are unable to capture the dynamic affinities between nodes and attributes. Moreover, existing co-embedding methods that learn the static embeddings of both nodes and attributes cannot be naturally utilized to obtain their dynamic embeddings for temporal attributed networks. To address these issues, we propose the dynamic co-embedding model for temporal attributed networks (DCTANs) based on the dynamic stochastic state-space framework. Our model captures the dynamics of a temporal attributed network by modeling the abstract belief states representing the condition of the nodes and attributes of current time step, and predicting the transitions between temporal abstract states of two successive time steps. Our model is able to learn embeddings for both nodes and attributes based on their belief states at each time step of the temporal attributed network, while the state transition tendency for predicting the future network can be tracked and the affinities between nodes and attributes can be preserved. Experimental results on real-world networks demonstrate that our model achieves substantial performance gains in several static and dynamic graph mining applications compared with the state-of-the-art static and dynamic models.

Multimodality Representation Learning, as a technique of learning to embed information from different modalities and their correlations, has achieved remarkable success on a variety of applications, such as Visual Question Answering (VQA), Natural Language for Visual Reasoning (NLVR), and Vision Language Retrieval (VLR). Among these applications, cross-modal interaction and complementary information from different modalities are crucial for advanced models to perform any multimodal task, e.g., understand, recognize, retrieve, or generate optimally. Researchers have proposed diverse methods to address these tasks. The different variants of transformer-based architectures performed extraordinarily on multiple modalities. This survey presents the comprehensive literature on the evolution and enhancement of deep learning multimodal architectures to deal with textual, visual and audio features for diverse cross-modal and modern multimodal tasks. This study summarizes the ( i ) recent task-specific deep learning methodologies, ( ii ) the pretraining types and multimodal pretraining objectives, ( iii ) from state-of-the-art pretrained multimodal approaches to unifying architectures, and ( iv ) multimodal task categories and possible future improvements that can be devised for better multimodal learning. Moreover, we prepare a dataset section for new researchers that covers most of the benchmarks for pretraining and finetuning. Finally, major challenges, gaps, and potential research topics are explored. A constantly-updated paperlist related to our survey is maintained at https://github.com/marslanm/multimodality-representation-learning .

Graph neural networks (GNNs) have gained significant attention for their impressive results on different graph-based tasks. The essential mechanism of GNNs is the message-passing framework, whereby node representations are aggregated from local neighborhoods. Recently, Transformer-based GNNs have been introduced to learn the long-range dependencies, enhancing performance. However, their quadratic computational complexity, due to the attention computation, has constrained their applicability on large-scale graphs. To address this issue, we propose MGIGNN ( M emorized G lobal I nformation G raph N eural N etwork), an innovative approach that leverages memorized global information to enhance existing GNNs in both transductive and inductive scenarios. Specifically, MGIGNN captures long-range dependencies by identifying and incorporating global similar nodes, which are defined as nodes exhibiting similar features, structural patterns and label information within a graph. To alleviate the computational overhead associated with computing embeddings for all nodes, we introduce an external memory module to facilitate the retrieval of embeddings and optimize performance on large graphs. To enhance the memory-efficiency, MGIGNN selectively retrieves global similar nodes from a small set of candidate nodes. These candidate nodes are selected from the training nodes based on a sparse node selection distribution with a Dirichlet prior. This selecting approach not only reduces the memory size required but also ensures efficient utilization of computational resources. Through comprehensive experiments conducted on ten widely-used and real-world datasets, including seven homogeneous datasets and three heterogeneous datasets, we demonstrate that our MGIGNN can generally improve the performance of existing GNNs on node classification tasks under both inductive and transductive settings.

Generating molecular graphs using deep graph generative models is a challenging task that involves optimizing a given target within an enormous search space while adhering to chemical valence rules. Despite promising results, existing models mainly focus on learning molecular graph structures at the individual level while ignoring inter-molecular relationships regarding molecular characterization features and molecular activity. This can lead to the generation of molecules that are unresponsive to their true neighbors possessing similar characterization features, resulting in a divergence between the learned generation distribution and the actual molecular distribution. In this paper, we propose a distribution preserving model, designed to maintain the inter-molecular relationships of the original distribution within the generated space. Specifically, the model operates on a student-teacher paradigm, where the teacher module learns the inter-molecular relationship dynamics of the original distribution, and imparts this knowledge to the student module, which is responsible for generating molecules. Experimental results on two benchmark datasets demonstrate that our model outperforms the state-of-the-art models in generating valid, novel and unique molecules. Moreover, our model is verified on preserving molecule distribution in the generation space.

The increasing volume of financial documents requires efficient summarization methods. This study investigates the use of natural language processing (NLP) to summarize financial narratives from annual reports in English, Spanish, and Greek. We employ T5 for English and mT5 for Greek and Spanish, generating structured summaries of firms’ yearly financial trends. Despite the challenges posed by diverse and unstructured reports, our approach effectively identifies key narrative elements, excluding financial tables and numerical data. In competition, our system significantly exceeded the baseline model, with placements varying by language and a weighted score distribution of 50% for English, 25% for Greek and 25% for Spanish, producing a composite score of 0.112.

This paper focuses on continual meta-learning, where few-shot tasks are heterogeneous and sequentially available. Recent works use a mixture model for meta-knowledge to deal with the heterogeneity. However, these methods suffer from parameter inefficiency caused by two reasons: (1) the underlying assumption of mutual exclusiveness among mixture components hinders sharing meta-knowledge across heterogeneous tasks. (2) they only allow increasing mixture components and cannot adaptively filter out redundant components. In this paper, we propose an Adaptive Compositional Continual Meta-Learning (ACML) algorithm, which employs a compositional premise to associate a task with a subset of mixture components, allowing meta-knowledge sharing among heterogeneous tasks. Moreover, to adaptively adjust the number of mixture components, we propose a component sparsification method based on evidential theory to filter out redundant components. Experimental results show ACML outperforms strong baselines, showing the effectiveness of our compositional meta-knowledge, and confirming that ACML can adaptively learn meta-knowledge.

Traditional methods of pre-training, fine-tuning, and ensembling often overlook essential relational data and task interconnections. To address this gap, our study presents a novel approach to harnessing this relational information via a relational graph-based model. We introduce Relational grAph Model ensemBLE model, abbreviated as RAMBLE. This model distinguishes itself by performing class label inference simultaneously across all data nodes and task nodes, employing the relational graph in a transductive manner. This fine-grained approach allows us to better comprehend and model the intricate interplay between data and tasks. Furthermore, we incorporate a novel variational information bottleneck-guided scheme for embedding fusion and aggregation. This innovative technique facilitates the creation of an informative fusion embedding, honing in on embeddings beneficial for the intended task while simultaneously filtering out potential noise-laden embeddings. Our theoretical analysis, grounded in information theory, confirms that the use of relational information for embedding fusion allows us to achieve higher upper and lower bounds on our target task's accuracy. We thoroughly assess our proposed model across eight diverse datasets, and the experimental results demonstrate the model's effective utilization of relational knowledge derived from all pre-trained models, thereby enhancing its performance on our target tasks.

Graph neural networks (GNNs) have become the de facto standard for representational learning in graphs, and have achieved state-of-the-art performance in many graph-related tasks; however, it has been shown that the expressive power of standard GNNs are equivalent maximally to 1-dimensional Weisfeiler-Lehman (1-WL) Test. Recently, there is a line of works aiming to enhance the expressive power of graph neural networks. One line of such works aim at developing K-hop message-passing GNNs where node representation is updated by aggregating information from not only direct neighbors but all neighbors within K-hop of the node. Another line of works leverages subgraph information to enhance the expressive power which is proven to be strictly more powerful than 1-WL test. In this work, we discuss the limitation of K-hop message-passing GNNs and propose substructure encoding function to uplift the expressive power of any K-hop message-passing GNN. We further inject contextualized substructure information to enhance the expressiveness of K-hop message-passing GNNs. Our method is provably more powerful than previous works on K-hop graph neural networks and 1-WL subgraph GNNs, which is a specific type of subgraph based GNN models, and not less powerful than 3-WL. Empirically, our proposed method set new state-of-the-art performance or achieves comparable performance for a variety of datasets. Our code is available at https://github.com/tianyao-aka/Expresive_K_hop_GNNs.

MicroRNAs (miRNAs) silence genes by binding to messenger RNAs, whereas long non-coding RNAs (lncRNAs) act as competitive endogenous RNAs (ceRNAs) that can relieve miRNA silencing effects and upregulate target gene expression. The ceRNA association between lncRNAs and miRNAs has been a research hotspot due to its medical importance, but it is challenging to verify experimentally. In this paper, we propose a novel deep learning scheme, i.e. sequence pre-training-based graph neural network (SPGNN), that combines pre-training and fine-tuning stages to predict lncRNA-miRNA associations from RNA sequences and the existing interactions represented as a graph. First, we utilize a sequence-to-vector technique to generate pre-trained embeddings based on the sequences of all RNAs during the pre-training stage. In the fine-tuning stage, we use Graph Neural Network to learn node representations from the heterogeneous graph constructed using lncRNA-miRNA association information. We evaluate our proposed scheme SPGNN on our newly collected animal lncRNA-miRNA association dataset and demonstrate that combining the $k$-mer technique and Doc2vec model for pre-training with the Simple Graph Convolution Network for fine-tuning is effective in predicting lncRNA-miRNA associations. Our approach outperforms state-of-the-art baselines across various evaluation metrics. We also conduct an ablation study and hyperparameter analysis to verify the effectiveness of each component and parameter of our scheme. The complete code and dataset are available on GitHub: https://github.com/zixwang/SPGNN.

Circular RNA (CircRNA) is a type of non-coding RNAs in which both ends are covalently linked. Researchers have demonstrated that many circRNAs can act as biomarkers of diseases. However, traditional experimental methods for circRNA-disease associations identification are labor-intensive. In this work, we propose a novel method based on the heterogeneous graph neural network and metapaths for circRNA-disease associations prediction termed as HMCDA. First, a heterogeneous graph consisting of circRNA-disease associations, circRNA-miRNA associations, miRNA-disease associations and disease-disease associations are constructed. Then, six metapaths are defined and generated according to the biomedical pathways. Afterwards, the entity content transformation, intra-metapath and inter-metapath aggregation are implemented to learn the embeddings of circRNA and disease entities. Finally, the learned embeddings are used to predict novel circRNA-disase associations. In particular, the result of extensive experiments demonstrates that HMCDA outperforms four state-of-the-art models in fivefold cross validation. In addition, our case study indicates that HMCDA has the ability to identify novel circRNA-disease associations.

In this paper, we study the task of cross-temporal snapshot alignment for dynamic networks. The goal of this task is to match equivalent nodes across temporal snapshots of a given dynamic network. Previous static network alignment methods ignore the non-stationary nature of networks, while existing dynamic counterparts focusing on two separate evolving networks lose sight of the problem of aligning two snapshots of the same dynamic network. To alleviate these issues, we propose a Cross-Temporal Snapshot Alignment model (CTSA), which maps nodes from different snapshots into the same semantic space and makes the equivalent nodes in the source and target snapshots to be aligned locate as closely as possible. Our CTSA model utilizes graph neural networks to embed nodes for each snapshot by aggregating the local structural information, and integrates the self-attention based encoders to model the dependencies in different snapshots over time. Additionally, to improve the alignment performance of the model, we contrive a novel positional embedding learning method, which takes into account both the ordering information of input representation sequences at each time step and the graph information of each network snapshot. Experimental results on real-world dynamic networks demonstrate that our model outperforms the state-of-the-art baselines.

Conversational Recommendation Systems (CRSs) aim to improve recommendation performance by utilizing information from a conversation session. A CRS first constructs questions and then asks users for their feedback in each conversation session to refine better recommendation lists to users. The key design of CRS is to construct proper questions and obtain users’ feedback in response to these questions so as to effectively capture user preferences. Many CRS works have been proposed; however, they suffer from defects when constructing questions for users to answer: (1) employing a dialogue policy agent for constructing questions is one of the most common choices in CRS, but it needs to be trained with a huge corpus, and (2) it is not appropriate that constructing questions from a single policy (e.g., a CRS only selects attributes that the user has interacted with) for all users with different preferences. To address these defects, we propose a novel CRS model, namely a Representation Fusion–based Conversational Recommendation model, where the whole conversation session is divided into two subsessions (i.e., Local Question Search subsession and Global Question Search subsession) and two different question search methods are proposed to construct questions in the corresponding subsessions without employing policy agents. In particular, in the Local Question Search subsession we adopt a novel graph mining method to find questions, where the paths in the graph between users and attributes can eliminate irrelevant attributes; in the Global Question Search subsession we propose to initialize user preference on items with the user and all item historical rating records and construct questions based on user’s preference. Then, we update the embeddings independently over the two subsessions according to user’s feedback and fuse the final embeddings from the two subsessions for the recommendation. Experiments on three real-world recommendation datasets demonstrate that our proposed method outperforms five state-of-the-art baselines.

In this article, we study the problem of dynamic personalized product search. Due to the data-sparsity problem in the real world, existing methods suffer from the challenge of data inefficiency. We address the challenge by proposing a Dynamic Bayesian Contrastive Predictive Coding model (DBCPC), which aims to capture the rich structured information behind search records to improve data efficiency. Our proposed DBCPC utilizes contrastive predictive learning to jointly learn dynamic embeddings with structure information of entities (i.e., users, products, and words). Specifically, our DBCPC employs structured prediction to tackle the intractability caused by non-linear output space and utilizes the time embedding technique to avoid designing different encoders each time in the Dynamic Bayesian models. In this way, our model jointly learns the underlying embeddings of entities (i.e., users, products, and words) via prediction tasks, which enables the embeddings to focus more on their general attributes and capture the general information during the preference evolution with time. For inferring the dynamic embeddings, we propose an inference algorithm combining the variational objective and the contrastive objectives. Experiments were conducted on an Amazon dataset and the experimental results show that our proposed DBCPC can learn the higher-quality embeddings and outperforms the state-of-the-art non-dynamic and dynamic models for product search.

Learning effective representations of entities and relations for knowledge graphs (KGs) is critical to the success of many multi-relational learning tasks. Existing methods based on graph neural networks learn a deterministic embedding function, which lacks sufficient flexibility to explore better choices when dealing with the imperfect and noisy KGs such as the scarce labeled nodes and noisy graph structure. To this end, we propose a novel multi-relational graph Gaussian Process network (GGPN), which aims to improve the flexibility of deterministic methods by simultaneously learning a family of embedding functions, i.e., a stochastic embedding function. Specifically, a Bayesian Gaussian Process (GP) is proposed to model the distribution of this stochastic function and the resulting representations are obtained by aggregating stochastic function values, i.e., messages, from neighboring entities. The two problems incurred when leveraging GP in GGPN are the proper choice of kernel function and the cubic computational complexity. To address the first problem, we further propose a novel kernel function that can explicitly take the diverse relations between each pair of entities into account and be adaptively learned in a data-driven way. We address the second problem by reformulating GP as a Bayesian linear model, resulting in a linear computational complexity. With these two solutions, our GGPN can be efficiently trained in an end-to-end manner. We evaluate our GGPN in link prediction and entity classification tasks, and the experimental results demonstrate the superiority of our method. Our code is available at https://github.com/sysu-gzchen/GGPN.

Parameter-Efficient Tuning (PETuning) methods have been deemed by many as the new paradigm for using pretrained language models (PLMs).By tuning just a fraction amount of parameters comparing to full model finetuning, PETuning methods claim to have achieved performance on par with or even better than finetuning.In this work, we take a step back and re-examine these PETuning methods by conducting the first comprehensive investigation into the training and evaluation of them.We found the problematic validation and testing practice in current studies, when accompanied by the instability nature of PETuning methods, has led to unreliable conclusions.When being compared under a truly fair evaluation protocol, PETuning cannot yield consistently competitive performance while finetuning remains to be the best-performing method in medium-and high-resource settings.We delve deeper into the cause of the instability and observed that the number of trainable parameters and training iterations are two main factors: reducing trainable parameters and prolonging training iterations may lead to higher stability in PETuning methods. 1

Personalized product search that provides users with customized search services is an important task for e-commerce platforms. This task remains a challenge when inferring users’ preferences from few records or even no records, which is also known as the few-shot or zero-shot learning problem. In this paper, we propose a Bayesian Online Meta-Learning Model (BOML), which transfers meta-knowledge, from the inference for other users’ preferences, to help to infer the current user’s interest behind her/his few or even no historical records. To extract meta-knowledge from various inference patterns, our model constructs a mixture of meta-knowledge and transfers the corresponding meta-knowledge to the specific user according to her/his records. Based on the meta-knowledge learned from other similar inferences, our proposed model searches a ranked list of products to meet users’ personalized query intents for those with few search records (i.e., few-shot learning problem) or even no search records (i.e., zero-shot learning problem). Under the records arriving sequentially setting, we propose an online variational inference algorithm to update meta-knowledge over time. Experimental results demonstrate that our proposed BOML outperforms state-of-the-art algorithms.

Network embedding is a technique that aims at inferring the low-dimensional representations of nodes in a semantic space. In this article, we study the problem of inferring the low-dimensional representations of both nodes and attributes for attributed networks in the same semantic space such that the affinity between a node and an attribute can be effectively measured. Intuitively, this problem can be addressed by simply utilizing existing variational auto-encoder (VAE) based network embedding algorithms. However, the variational posterior distribution in previous VAE based network embedding algorithms is often assumed and restricted to be a mean-field Gaussian distribution or other simple distribution families, which results in poor inference of the embeddings. To alleviate the above defect, we propose a novel VAE-based co-embedding method for attributed network, F-CAN, where posterior distributions are flexible, complex, and scalable distributions constructed through the normalizing flow. We evaluate our proposed models on a number of network tasks with several benchmark datasets. Experimental results demonstrate that there are clear improvements in the qualities of embeddings generated by our model to the state-of-the-art attributed network embedding methods.

Network-based information has been widely explored and exploited in the information retrieval literature. Attributed networks, consisting of nodes, edges as well as attributes describing properties of nodes, are a basic type of network-based data, and are especially useful for many applications. Examples include user profiling in social networks and item recommendation in user-item purchase networks. Learning useful and expressive representations of entities in attributed networks can provide more effective building blocks to down-stream network-based tasks such as link prediction and attribute inference. Practically, input features of attributed networks are normalized as unit directional vectors. However, most network embedding techniques ignore thesphericalnature of inputs and focus on learning representations in a Gaussian or Euclidean space, which, we hypothesize, might lead to less effective representations. To obtain more effective representations of attributed networks, we investigate the problem of mapping an attributed network with unit normalized directional features into a non-Gaussian and non-Euclidean space. Specifically, we propose a hyperspherical variational co-embedding for attributed networks (HCAN), which is based on generalized variational auto-encoders for heterogeneous data with multiple types of entities. HCAN jointly learns latent embeddings for both nodes and attributes in a unified hyperspherical space such that the affinities between nodes and attributes can be captured effectively. We argue that this is a crucial feature in many real-world applications of attributed networks. Previous Gaussian network embedding algorithms break the assumption of uninformative prior, which leads to unstable results and poor performance. In contrast, HCAN embeds nodes and attributes as von Mises-Fisher distributions, and allows one to capture the uncertainty of the inferred representations. Experimental results on eight datasets show that HCAN yields better performance in a number of applications compared with nine state-of-the-art baselines.

In this article, we study the task of user profiling in question answering communities (QACs). Previous user profiling algorithms suffer from a number of defects: they regard users and words as atomic units, leading to the mismatch between them; they are designed for other applications but not for QACs; and some semantic profiling algorithms do not co-embed users and words, leading to making the affinity measurement between them difficult. To improve the profiling performance, we propose a neural Flow-based Constrained Co-embedding Model, abbreviated as FCCM. FCCM jointly co-embeds the vector representations of both users and words in QACs such that the affinities between them can be semantically measured. Specifically, FCCM extends the standard variational auto-encoder model to enforce the inferred embeddings of users and words subject to the voting constraint, i.e., given a question and the users who answer this question in the community, representations of the users whose answers receive more votes are closer to the representations of the words associated with these answers, compared with representations of whose receiving fewer votes. In addition, FCCM integrates normalizing flow into the variational auto-encoder framework to avoid the assumption that the distributions of the embeddings are Gaussian, making the inferred embeddings fit the real distributions of the data better. Experimental results on a Chinese Zhihu question answering dataset demonstrate the effectiveness of our proposed FCCM model for the task of user profiling in QACs.

The task of personalized product search aims at retrieving a ranked list of products given a user’s input query and his/her purchase history. To address this task, we propose the PSAM model, a Personalized, Sequential, Attentive and Metric-aware (PSAM) model, that learns the semantic representations of three different categories of entities, i.e., users, queries, and products, based on user sequential purchase historical data and the corresponding sequential queries. Specifically, a query-based attentive LSTM (QA-LSTM) model and an attention mechanism are designed to infer users dynamic embeddings, which is able to capture their short-term and long-term preferences. To obtain more fine-grained embeddings of the three categories of entities, a metric-aware objective is deployed in our model to force the inferred embeddings subject to the triangle inequality, which is a more realistic distance measurement for product search. Experiments conducted on four benchmark datasets show that our PSAM model significantly outperforms the state-of-the-art product search baselines in terms of effectiveness by up to 50.9% improvement under NDCG@20. Our visualization experiments further illustrate that the learned product embeddings are able to distinguish different types of products.

Mobile advertising has undoubtedly become one of the fastest-growing industries in the world. The influx of capital attracts increasing fraudsters to defraud money from advertisers. Fraudsters can leverage many techniques, where bots install fraud is the most difficult to detect due to its ability to emulate normal users by implementing sophisticated behavioral patterns to evade from detection rules defined by human experts. Therefore, we proposed BotSpot 1 for bots install fraud detection previously. However, there are some drawbacks in BotSpot, such as the sparsity of the devices’ neighbors, weak interactive information of leaf nodes, and noisy labels. In this work, we propose BotSpot++ to improve these drawbacks: (1) for the sparsity of the devices’ neighbors, we propose to construct a super device node to enrich the graph structure and information flow utilizing domain knowledge and a clustering algorithm; (2) for the weak interactive information, we propose to incorporate a self-attention mechanism to enhance the interaction of various leaf nodes; and (3) for the noisy labels, we apply a label smoothing mechanism to alleviate it. Comprehensive experimental results show that BotSpot++ yields the best performance compared with six state-of-the-art baselines. Furthermore, we deploy our model to the advertising platform of Mobvista, 2 a leading global mobile advertising company. The online experiments also demonstrate the effectiveness of our proposed method.

Deep quantization methods have shown high efficiency on large-scale image retrieval. However, current models heavily rely on ground-truth information, hindering the application of quantization in label-hungry scenarios. A more realistic demand is to learn from inexhaustible uploaded images that are associated with informal tags provided by amateur users. Though such sketchy tags do not obviously reveal the labels, they actually contain useful semantic information for supervising deep quantization. To this end, we propose Weakly-Supervised Deep Hyperspherical Quantization (WSDHQ), which is the first work to learn deep quantization from weakly tagged images. Specifically, 1) we use word embeddings to represent the tags and enhance their semantic information based on a tag correlation graph. 2) To better preserve semantic information in quantization codes and reduce quantization error, we jointly learn semantics-preserving embeddings and supervised quantizer on hypersphere by employing a well-designed fusion layer and tailor-made loss functions. Extensive experiments show that WSDHQ can achieve state-of-art performance in weakly-supervised compact coding.

Gaussian Process (GP) offers a principled non-parametric framework for learning stochastic functions. The generalization capability of GPs depends heavily on the kernel function, which implicitly imposes the smoothness assumptions of the data. However, common feature-based kernel functions are inefficient to model the relational data, where the smoothness assumptions implied by the kernels are violated. To model the complex and non-differentiable functions over relational data, we propose a novel Graph Convolutional Kernel, which enables to incorporate relational structures to feature-based kernels to capture the statistical structure of data. To validate the effectiveness of proposed kernel function in modeling relational data, we introduce GP models with Graph Convolutional Kernel in two relational learning settings, i.e., unsupervised settings of link prediction and semi-supervised settings of object classification. The parameters of our GP models are optimized through the scalable variational inducing point method. However, the highly structured likelihood objective requires densely sampling from variational distributions, which is costly and makes its optimization challenging in the unsupervised settings. To tackle this challenge, we propose a Local Neighbor Sampling technique with a provable more efficient computational complexity. Experimental results on real-world datasets demonstrate that our model achieves state-of-the-art performance in two relational learning tasks.

Gaussian Processes (GPs) define distributions over functions and their generalization capabilities depend heavily on the choice of kernels. In this paper, we propose a novel structure-aware random Fourier (SRF) kernel for GPs that brings several benefits when modeling graph-structured data. First, SRF kernel is defined with a spectral distribution based on the Fourier duality given by the Bochner's theorem, transforming the kernel learning problem to a distribution inference problem. Second, SRF kernel admits a random Fourier feature formulation that makes the kernel scalable for optimization. Third, SRF kernel enables to leverage geometric structures by taking subgraphs as inputs. To effectively optimize GPs with SRF kernel, we develop a variational EM algorithm, which alternates between an inference procedure (E-step) and a learning procedure (M-step). Experimental results on five real-world datasets show that our model can achieve state-of-the-art performance in two typical graph learning tasks, i.e., object classification and link prediction.

Conventional meta-learning considers a set of tasks from a stationary distribution. In contrast, this paper focuses on a more complex online setting, where tasks arrive sequentially and follow a non-stationary distribution. Accordingly, we propose a Variational Continual Bayesian Meta-Learning (VC-BML) algorithm. VC-BML maintains a Dynamic Gaussian Mixture Model for meta-parameters, with the number of component distributions determined by a Chinese Restaurant Process. Dynamic mixtures at the meta-parameter level increase the capability to adapt to diverse tasks due to a larger parameter space, alleviating the negative knowledge transfer problem. To infer posteriors of model parameters, compared to the previously used point estimation method, we develop a more robust posterior approximation method -- structured variational inference for the sake of avoiding forgetting knowledge. Experiments on tasks from non-stationary distributions show that VC-BML is superior in transferring knowledge among diverse tasks and alleviating catastrophic forgetting in an online setting.

Representation learning for temporal knowledge graphs has attracted increasing attention in recent years. In this paper, we study the problem of learning dynamic embeddings for temporal knowledge graphs. We address this problem by proposing a Dynamic Bayesian Knowledge Graphs Embedding model (DBKGE), which is able to dynamically track the semantic representations of entities over time in a joint metric space and make predictions for the future. Unlike other temporal knowledge graph embedding methods, DBKGE is a novel probabilistic representation learning method that aims at inferring dynamic embeddings of entities in a streaming scenario. To obtain high-quality embeddings and model their uncertainty, our DBKGE embeds entities with means and variances of Gaussian distributions. Based on amortized inference, an online inference algorithm is proposed to jointly learn the latent representations of entities and smooth their changes across time. Experiments on Yago and Wiki datasets demonstrate that our proposed algorithm outperforms the state-of-the-art static and temporal knowledge graph embedding models.

Mobile advertising has become inarguably one of the fastest growing industries all over the world. The influx of capital attracts increasing fraudsters to defraud money from advertisers. There are many tricks a fraudster can leverage, among which bot install fraud is undoubtedly the most insidious one due to its ability to implement sophisticated behavioral patterns and emulate normal users, so as to evade from detection rules defined by human experts. In this work, we propose an anti-fraud method based on heterogeneous graph that incorporates both local context and global context via graph neural networks (GNN) and gradient boosting classifier to detect bot fraud installs at Mobvista, a leading global mobile advertising company. Offline evaluations in two datasets show the proposed method outperforms all the competitive baseline methods by at least 2.2% in the first dataset and 5.75% in the second dataset given the evaluation metric [email protected]% Precision. Furthermore, we deploy our method to tackle million-scale data daily at Mobvista. The online performance also shows that the proposed methods consistently detect more bots than other baseline methods.

The field of recommender systems has rapidly evolved over the last few years, with significant advances made due to the in-flux of deep learning techniques. However, as a result of this rapid progress, escalating barriers-to-entry for new researchers is emerging. In particular, state-of-the-art approaches have fragmented into a large number of code-bases, often requiring different input formats, pre-processing stages and evaluating with different metric packages. Hence, it is time-consuming for new researchers to reach the point of having both an effective baseline set and a sound comparative environment. As a step towards elevating this problem, we have developed BETA-Rec, an open source project for Building, Evaluating and Tuning Automated Recommender Systems. BETA-Rec aims to provide a practical data toolkit for building end-to-end recommendation systems in a standardized way. It provides means for dataset preparation and splitting using common strategies, a generalized model engine for implementing recommender models using Pytorch with 9 models available out-of-the-box, as well as a unified training, validation, tuning and testing pipeline. Furthermore, BETA-Rec is designed to be both modular and extensible, enabling new models to be quickly added to the framework. It is deployable in a wide range of environments via pre-built docker containers and supports distributed parameter tuning using Ray. In this demo, we will illustrate the deployment and use of BETA-Rec for researchers and practitioners on a number of standard recommendation datasets. The source code of the project is available at github: https://github.com/beta-team/beta-recsys.

Previous embedding methods for attributed networks aim at learning low-dimensional vector representations only for nodes but not for both nodes and attributes, resulting in the fact that node embeddings cannot be directly used to recover the correlations between nodes and attributes. However, capturing such correlations by embeddings is of great importance for many real-world applications, such as attribute inference and user profiling. Moreover, in real-world scenarios, many attributed networks evolve over time, with their nodes, links, and attributes changing from time to time. In this article, we study the problem of jointly learning low-dimensional representations of both nodes and attributes for static and dynamic attributed networks. To address this problem, we propose a Co-embedding model for Static Attributed Networks (CSAN), which jointly learns low-dimensional representations of both attributes and nodes in the same semantic space such that their affinities can be effectively captured and measured, and a Co-embedding model for Dynamic Attributed Networks (CDAN) to dynamically track low-dimensional representations of nodes and attributes over time. To obtain effective embeddings, both our co-embedding models, CSAN and CDAN, embed each node and attribute with means and variances of Gaussian distributions via variational auto-encoders. Our CDAN model formulates the dynamic changes of a dynamic attributed network by aggregating perturbation features from the nodes’ local neighborhoods as well as attributes’ associations such that the evolving patterns of the given network can be tracked. Experimental results on real-world networks demonstrate that our proposed embedding models outperform state-of-the-art non-dynamic and dynamic embedding models.

In this paper, we study the problem of dynamic user profiling in the context of streams of short texts. Previous work on user profiling works with long documents, do not consider collaborative information, and do not diversify the keywords for profiling users’ interests. In contrast, we address the problem by proposing a user profiling algorithm (UPA), which consists of two models: the proposed collaborative interest tracking topic model (CITM) and the proposed streaming keyword diversification model (SKDM). UPA first utilizes CITM to collaboratively track each user’s and his followees’ dynamic interest distributions in the context of streams of short texts, and then utilizes SKDM to obtain top-k relevant and diversified keywords to profile users’ interests at a specific point in time. Experiments were conducted on a Twitter dataset and we found that UPA outperforms state-of-the-art non-dynamic and dynamic user profiling algorithms.

In this paper, we propose a Bayesian Deep Collaborative Matrix Factorization (BDCMF) algorithm for collaborative filtering (CF). BDCMF is a novel Bayesian deep generative model that learns user and item latent vectors from users’ social interactions, contents of items as the auxiliary information and user-item rating (feedback) matrix. It alleviates the problem of matrix sparsity by incorporating items’ auxiliary and users’ social information into the model. It can learn more robust and dense latent representations by integrating deep learning into Bayesian probabilistic framework. As being one of deep generative models, it has both non-linearity and Bayesian nature. Additionally, in BDCMF, we derive an efficient EM-style point estimation algorithm for parameter learning. To further improve recommendation performance, we also derive a full Bayesian posterior estimation algorithm for inference. Experiments conducted on two sparse datasets show that BDCMF can significantly outperform the state-of-the-art CF methods.

In modern e-commerce, the temporal order behind users’ transactions implies the importance of exploiting the transition dependency among items for better inferring what a user prefers to interact in “near future”. The types of interaction among items are usually divided into individual-level interaction that can stand out the transition order between a pair of items, or union-level relation between a set of items and single one. However, most of existing work only captures one of them from a single view, especially on modeling the individual-level interaction. In this paper, we propose a Multi-order Attentive Ranking Model (MARank) to unify both individual- and union-level item interaction into preference inference model from multiple views. The idea is to represent user’s short-term preference by embedding user himself and a set of present items into multi-order features from intermedia hidden status of a deep neural network. With the help of attention mechanism, we can obtain a unified embedding to keep the individual-level interactions with a linear combination of mapped items’ features. Then, we feed the aggregated embedding to a designed residual neural network to capture union-level interaction. Thorough experiments are conducted to show the features of MARank under various component settings. Furthermore experimental results on several public datasets show that MARank significantly outperforms the state-of-the-art baselines on different evaluation metrics. The source code can be found at https://github.com/voladorlu/MARank.

In this paper, we study the problem of user profiling in question answering communities. We address the problem by proposing a constrained co-embedding model (CCEM). CCEM jointly infers the embeddings of both users and words in question answering communities such that the similarities between users and words can be semantically measured. Our CCEM works with constraints which enforce the inferred embeddings of users and words subject to this criteria: given a question in the community, embeddings of users whose answers receive more votes are closer to the embeddings of the words occurring in these answers, compared to the embeddings of those whose answers receive less votes. Experiments on a Chinese dataset, Zhihu dataset, demonstrate that our proposed co-embedding algorithm outperforms state-of-the-art methods in the task of user profiling.

In this paper, we provide a unified learning algorithm, dynamic collaborative recurrent learning, DCRL, of two directions of recommendations: temporal recommendations focusing on tracking the evolution of users' long-term preference and sequential recommendations focusing on capturing short-term preferences given a short time window. Our DCRL builds based on RNN and Sate Space Model (SSM), and thus it is not only able to collaboratively capture users' short-term and long-term preferences as in sequential recommendations, but also can dynamically track the evolution of users' long-term preferences as in temporal recommendations in a unified framework. In addition, we introduce two smoothing and filtering scalable inference algorithms for DCRL's offline and online learning, respectively, based on amortized variational inference, allowing us to effectively train the model jointly over all time. Experiments demonstrate DCRL outperforms the temporal and sequential recommender models, and does capture users' short-term preferences and track the evolution of long-term preferences.

In this paper, we study the problem of personalized product search under streaming scenarios. We address the problem by proposing a Dynamic Bayesian Metric Learning model, abbreviated as DBML, which can collaboratively track the evolutions of latent semantic representations of different categories of entities (i.e., users, products and words) over time in a joint metric space. In particular, unlike previous work using inner-product metric to model the affinities between entities, our DBML is a novel probabilistic metric learning approach that is able to avoid the contradicts, keep the triangle inequality in the latent space, and correctly utilize implicit feedbacks. For inferring dynamic embeddings of the entities, we propose a scalable online inference algorithm, which can jointly learn the latent representations of entities and smooth their changes across time, based on amortized inference. The inferred dynamic semantic representations of entities collaboratively inferred in a unified form by our DBML can benefit not only for improving personalized product search, but also for capturing the affinities between users, products and words. Experimental results on large datasets over a number of applications demonstrate that our DBML outperforms the state-of-the-art algorithms, and can effectively capture the evolutions of semantic representations of different categories of entities over time.

Session-based recommendation is a challenging problem due to the inherent uncertainty of user behavior and the limited historical click information. Latent factors and the complex dependencies within the user’s current session have an important impact on the user's main intention, but the existing methods do not explicitly consider this point. In this paper, we propose a novel model, Interest Shift and Latent Factors Combination Model (ISLF), which can capture the user's main intention by taking into account the user’s interest shift (i.e. long-term and short-term interest) and latent factors simultaneously. In addition, we experimentally give an explicit explanation of this combination in our ISLF. Our experimental results on three benchmark datasets show that our model achieves state-of-the-art performance on all test datasets.

Deep generative models (DGMs) have achieved remarkable advances. Semi-supervised variational auto-encoders (SVAE) as a classical DGM offers a principled framework to effective generalize from small labelled data to large unlabelled ones, but it is difficult to incorporate rich unstructured relationships within the multiple heterogeneous entities. In this paper, to deal with the problem, we present a semi-supervised co-embedding model for attributed networks (SCAN) based on the generalized SVAE for the heterogeneous data, which collaboratively learns low- dimensional vector representations of both nodes and attributes for partially labelled attributed networks semi-supervisedly. The node and attribute embeddings obtained in a unified manner by our SCAN can benefit not only for capturing the proximities between nodes but also the affinities between nodes and attributes. Moreover, our model also trains a discriminative network to learn the label predictive distribution of nodes. Experimental results on real-world networks demonstrate that our model yields excellent performance in a number of applications such as attribute inference, user profiling and node classification compared to the state-of-the-art baselines.

Existing embedding methods for attributed networks aim at learning low-dimensional vector representations for nodes only but not for both nodes and attributes, resulting in the fact that they cannot capture the affinities between nodes and attributes. However, capturing such affinities is of great importance to the success of many real-world attributed network applications, such as attribute inference and user profiling. Accordingly, in this paper, we introduce a Co-embedding model for Attributed Networks (CAN), which learns low-dimensional representations of both attributes and nodes in the same semantic space such that the affinities between them can be effectively captured and measured. To obtain high-quality embeddings, we propose a variational auto-encoder that embeds each node and attribute with means and variances of Gaussian distributions. Experimental results on real-world networks demonstrate that our model yields excellent performance in a number of applications compared with state-of-the-art techniques.

Sources in computer-based collaborative systems such as webpages can help employees to connect and cooperate with each other. It is natural to enable the systems to look not only for documents but also for experts. In this paper, we study the problem of expert retrieval in enterprise corpora: given a topic, also known as query containing a set of words, identify a rank list of candidate experts who have expertise on the topic. To tackle the problem, we propose an unsupervised semantic two-player minimax game, i.e., our unsupervised semantic generative adversarial networks (USGAN). Unlike almost all the previous generative adversarial networks-based algorithms that require ground truth training data, our USGAN is an unsupervised semantic expert retrieval algorithm that consists of a discriminative network and a generative network aiming at capturing the representations of words and experts in an unsupervised way. Candidates that have similar semantic representations to that of the topic are retrieved as relevant to the topic. Our USGAN would provide inspiration on how to extend the standard GAN and its variants by unsupervised ways to address other retrieval tasks where labelled data are missing. Experimental results on public datasets validate the effectiveness of the proposed expert retrieval algorithm.

In this paper, we study the problem of recommending personalized items to users given their sequential behaviors. Most sequential recommendation models only capture a user's short-term preference in a short session, and neglect his general (unchanged over time) and long-term preferences. Besides, they are all based on deterministic neural networks, and consider users' latent preferences as point vectors in a low-dimensional continuous space. However, in real world, the evolutions of users' preferences are full of uncertainties. We address this problem by proposing a hierarchical neural variational model (HNVM). HNVM models users' three preferences: general, long-term and short-term preferences through an unified hierarchical deep generative process. HNVM is a hierarchical recurrent neural network that enables it to capture both user's long-term and short-term preferences. Experiments on two public datasets demonstrate that HNVM outperforms state-of-the-art sequential recommendation methods.

Stance detection has gained increasing interest from the research community due to its importance for fake news detection. The goal of stance detection is to categorize an overall position of a subject towards an object into one of the four classes: agree, disagree, discuss, and unrelated. One of the major problems faced by current machine learning models used for stance detection is caused by a severe class imbalance among these classes. Hence, most models fail to correctly classify instances that fall into minority classes. In this paper, we address this problem by proposing a hierarchical representation of these classes, which combines the agree, disagree, and discuss classes under a new related class. Further, we propose a two-layer neural network that learns from this hierarchical representation and controls the error propagation between the two layers using the Maximum Mean Discrepancy regularizer. Compared with conventional four-way classifiers, this model has two advantages: (1) the hierarchical architecture mitigates the class imbalance problem; (2) the regularization makes the model to better discern between the related and unrelated stances. An extensive experimentation demonstrates state-of-the-art accuracy performance of the proposed model for stance detection.

Social media platforms are a plethora of misinformation and its potential negative influence on the public is a growing concern. This concern has drawn the attention of the research community on developing mechanisms to detect misinformation. The task of misinformation detection consists of classifying whether a claim is True or False. Most research concentrates on developing machine learning models, such as neural networks, that outputs a single value in order to predict the veracity of a claim. One of the major problem faced by these models is the inability of representing the uncertainty of the prediction, which is due incomplete or finite available information about the claim being examined. We address this problem by proposing a Bayesian deep learning model. The Bayesian model outputs a distribution used to represent both the prediction and its uncertainty. In addition to the claim content, we also encode auxiliary information given by people's replies to the claim. First, the model encodes a claim to be verified, and generate a prior belief distribution from which we sample a latent variable. Second, the model encodes all the people's replies to the claim in a temporal order through a Long Short Term Memory network in order to summarize their content. This summary is then used to update the prior belief generating the posterior belief. Moreover, in order to train this model, we develop a Stochastic Gradient Variational Bayes algorithm to approximate the analytically intractable posterior distribution. Experiments conducted on two public datasets demonstrate that our model outperforms the state-of-the-art detection models.

In this paper, we aim at tackling the problem of user clustering in the context of their published short text streams. Clustering users by short text streams is more challenging than in the case of long documents associated with them as it is difficult to track users' dynamic interests in streaming sparse data. To obtain better user clustering performance, we propose two user collaborative interest tracking models that aim at tracking changes of each user's dynamic topic distributions in collaboration with their followees' dynamic topic distributions, based both on the content of current short texts and the previously estimated distributions. Our models can be either short-term or long-term dependency topic models. Short-term dependency model collaboratively tracks users' interests based on users' topic distributions at the previous time period only, whereas long-term dependency model collaboratively tracks users' interests based on users' topic distributions at multiple time periods in the past. We also propose two collapsed Gibbs sampling algorithms for collaboratively inferring users' dynamic interests for their clustering in our short-term and long-term dependency topic models, respectively. We evaluate our proposed models via a benchmark dataset consisting of Twitter users and their tweets. Experimental results validate the effectiveness of our proposed models that integrate both users' and their collaborative interests for user clustering by short text streams.

In this paper, we aim at tackling the problem of dynamic user profiling in the context of streams of short texts. Profiling users' expertise in such context is more challenging than in the case of long documents in static collection as it is difficult to track users' dynamic expertise in streaming sparse data. To obtain better profiling performance, we propose a streaming profiling algorithm (SPA). SPA first utilizes the proposed user expertise tracking topic model (UET) to track the changes of users' dynamic expertise and then utilizes the proposed streaming keyword diversification algorithm (SKDA) to produce top-k diversified keywords for profiling users' dynamic expertise at a specific point in time. Experimental results validate the effectiveness of the proposed algorithms.

The task of dialogue generation aims to automatically provide responses given previous utterances. Tracking dialogue states is an important ingredient in dialogue generation for estimating users' intention. However, the expensive nature of state labeling and the weak interpretability make the dialogue state tracking a challenging problem for both task-oriented and non-task-oriented dialogue generation: For generating responses in task-oriented dialogues, state tracking is usually learned from manually annotated corpora, where the human annotation is expensive for training; for generating responses in non-task-oriented dialogues, most of existing work neglects the explicit state tracking due to the unlimited number of dialogue states.

In this paper, we study the problem of dynamic user profiling in Twitter. We address the problem by proposing a dynamic user and word embedding model (DUWE), a scalable black-box variational inference algorithm, and a streaming keyword diversification model (SKDM). DUWE dynamically tracks the semantic representations of users and words over time and models their embeddings in the same space so that their similarities can be effectively measured. Our inference algorithm works with a convex objective function that ensures the robustness of the learnt embeddings. SKDM aims at retrieving top-K relevant and diversified keywords to profile users' dynamic interests. Experiments on a Twitter dataset demonstrate that our proposed embedding algorithms outperform state-of-the-art non-dynamic and dynamic embedding and topic models.

We address the task of fusing ranked lists of documents that are retrieved in response to a query. Past work on this task of rank aggregation often assumes that documents in the lists being fused are independent and that only the documents that are ranked high in many lists are likely to be relevant to a given topic. We propose manifold learning aggregation approaches, ManX and v-ManX, that build on the cluster hypothesis and exploit inter-document similarity information. ManX regularizes document fusion scores, so that documents that appear to be similar within a manifold, receive similar scores, whereas v-ManX first generates virtual adversarial documents and then regularizes the fusion scores of both original and virtual adversarial documents. Since aggregation methods built on the cluster hypothesis are computationally expensive, we adopt an optimization method that uses the top-k documents as anchors and considerably reduces the computational complexity of manifold-based methods, resulting in two efficient aggregation approaches, a-ManX and a-v-ManX. We assess the proposed approaches experimentally and show that they significantly outperform the state-of-the-art aggregation approaches, while a-ManX and a-v-ManX run faster than ManX, v-ManX, respectively.

In this paper, we study the problem of user clustering in the context of their published short text streams. Clustering users by short text streams is more challenging than in the case of long documents associated with them as it is difficult to track users' dynamic interests in streaming sparse data. To obtain better user clustering performance, we propose a user collaborative interest tracking model (UCIT) that aims at tracking changes of each user's dynamic topic distributions in collaboration with their followees', based both on the content of current short texts and the previously estimated distributions. We evaluate our proposed method via a benchmark dataset consisting of Twitter users and their tweets. Experimental results validate the effectiveness of our proposed UCIT model that integrates both users' and their collaborative interests for user clustering by short text streams.

A recommendation is called explainable if it not only predicts a numerical rating for an item, but also generates explanations for users' preferences. Most existing methods for explainable recommendation apply topic models to analyze user reviews to provide descriptions along with the recommendations they produce. So far, such methods have neglected user opinions and influences from social relations as a source of information for recommendations, even though these are known to improve the rating prediction.

We aim at summarizing answers in community question-answering (CQA). While most previous work focuses on factoid question-answering, we focus on the non-factoid question-answering. Unlike factoid CQA, non-factoid question-answering usually requires passages as answers. The shortness, sparsity and diversity of answers form interesting challenges for summarization. To tackle these challenges, we propose a sparse coding-based summarization strategy that includes three core ingredients: short document expansion, sentence vectorization, and a sparse-coding optimization framework. Specifically, we extend each answer in a question-answering thread to a more comprehensive representation via entity linking and sentence ranking strategies. From answers extended in this manner, each sentence is represented as a feature vector trained from a short text convolutional neural network model. We then use these sentence representations to estimate the saliency of candidate sentences via a sparse-coding framework that jointly considers candidate sentences and Wikipedia sentences as reconstruction items. Given the saliency vectors for all candidate sentences, we extract sentences to generate an answer summary based on a maximal marginal relevance algorithm. Experimental results on a benchmark data collection confirm the effectiveness of our proposed method in answer summarization of non-factoid CQA, and moreover, its significant improvement compared to state-of-the-art baselines in terms of ROUGE metrics.

User clustering has been studied from different angles. In order to identify shared interests, behavior-based methods consider similar browsing or search patterns of users, whereas content-based methods use information from the contents of the documents visited by the users. So far, content-based user clustering has mostly focused on static sets of relatively long documents. Given the dynamic nature of social media, there is a need to dynamically cluster users in the context of streams of short texts. User clustering in this setting is more challenging than in the case of long documents, as it is difficult to capture the users’ dynamic topic distributions in sparse data settings. To address this problem, we propose a dynamic user clustering topic model (UCT). UCT adaptively tracks changes of each user’s time-varying topic distributions based both on the short texts the user posts during a given time period and on previously estimated distributions. To infer changes, we propose a Gibbs sampling algorithm where a set of word pairs from each user is constructed for sampling. UCT can be used in two ways: (1) as a short-term dependency model that infers a user’s current topic distribution based on the user’s topic distributions during the previous time period only, and (2) as a long-term dependency model that infers a user’s current topic distributions based on the user’s topic distributions during multiple time periods in the past. The clustering results are explainable and human-understandable, in contrast to many other clustering algorithms. For evaluation purposes, we work with a dataset consisting of users and tweets from each user. Experimental results demonstrate the effectiveness of our proposed short-term and long-term dependency user clustering models compared to state-of-the-art baselines.

We consider the problem of search result diversification for streams of short texts. Diversifying search results in short text streams is more challenging than in the case of long documents, as it is difficult to capture the latent topics of short documents. To capture the changes of topics and the probabilities of documents for a given query at a specific time in a short text stream, we propose a dynamic Dirichlet multinomial mixture topic model, called D2M3, as well as a Gibbs sampling algorithm for the inference. We also propose a streaming diversification algorithm, SDA, that integrates the information captured by D2M3 with our proposed modified version of the PM-2 (Proportionality-based diversification Method -- second version) diversification algorithm. We conduct experiments on a Twitter dataset and find that SDA statistically significantly outperforms state-of-the-art non-streaming retrieval methods, plain streaming retrieval methods, as well as streaming diversification methods that use other dynamic topic models.

Clustering technology has found numerous applications in mining textual data. It was shown to enhance the performance of retrieval systems in various different ways, such as identifying different query aspects in search result diversification, improving smoothing in the context of language modeling, matching queries with documents in a latent topic space in ad-hoc retrieval, summarizing documents etc. The vast majority of clustering methods have been developed under the assumption of a static corpus of long (and hence textually rich) documents. Little attention has been given to streaming corpora of short text, which is the predominant type of data in Web 2.0 applications, such as social media, forums, and blogs. In this paper, we consider the problem of dynamically clustering a streaming corpus of short documents. The short length of documents makes the inference of the latent topic distribution challenging, while the temporal dynamics of streams allow topic distributions to change over time. To tackle these two challenges we propose a new dynamic clustering topic model - DCT - that enables tracking the time-varying distributions of topics over documents and words over topics. DCT models temporal dynamics by a short-term or long-term dependency model over sequential data, and overcomes the difficulty of handling short text by assigning a single topic to each short document and using the distributions inferred at a certain point in time as priors for the next inference, allowing the aggregation of information. At the same time, taking a Bayesian approach allows evidence obtained from new streaming documents to change the topic distribution. Our experimental results demonstrate that the proposed clustering algorithm outperforms state-of-the-art dynamic and non-dynamic clustering topic models in terms of perplexity and when integrated in a cluster-based query likelihood model it also outperforms state-of-the-art models in terms of retrieval quality.

User clustering has been studied from different angles: behavior-based, to identify similar browsing or search patterns, and content-based, to identify shared interests. Once user clusters have been found, they can be used for recommendation and personalization. So far, content-based user clustering has mostly focused on static sets of relatively long documents. Given the dynamic nature of social media, there is a need to dynamically cluster users in the context of short text streams. User clustering in this setting is more challenging than in the case of long documents as it is difficult to capture the users' dynamic topic distributions in sparse data settings. To address this problem, we propose a dynamic user clustering topic model (or UCT for short). UCT adaptively tracks changes of each user's time-varying topic distribution based both on the short texts the user posts during a given time period and on the previously estimated distribution. To infer changes, we propose a Gibbs sampling algorithm where a set of word-pairs from each user is constructed for sampling. The clustering results are explainable and human-understandable, in contrast to many other clustering algorithms. For evaluation purposes, we work with a dataset consisting of users and tweets from each user. Experimental results demonstrate the effectiveness of our proposed clustering model compared to state-of-the-art baselines.

Query auto completion (QAC) methods recommend queries to search engine users when they start entering a query. Current QAC methods mostly rank query completions based on their past popularity, i.e., on the number of times they have previously been submitted as a query. However, query popularity changes over time and may vary drastically across users. Accordingly, the ranking of query completions should be adjusted. Previous time-sensitive and user-specific QAC methods have been developed separately, yielding significant improvements over methods that are neither time-sensitive nor personalized. We propose a hybrid QAC method that is both time-sensitive and personalized. We extend it to handle long-tail prefixes, which we achieve by assigning optimal weights to the contribution from time-sensitivity and personalization. Using real-world search log datasets, we return top$N$query suggestions ranked by predicted popularity as estimated from popularity trends and cyclic popularity behavior; we rerank them by integrating similarities to a user's previous queries (both in the current session and in previous sessions). Our method outperforms state-of-the-art time-sensitive QAC baselines, achieving total improvements of between 3 and 7 percent in terms of mean reciprocal rank (MRR). After optimizing the weights, our extended model achieves MRR improvements of between 4 and 8 percent.

This paper is concerned with the problem of personalized diversification of search results, with the goal of enhancing the performance of both plain diversification and plain personalization algorithms. In previous work, the problem has mainly been tackled by means of unsupervised learning. To further enhance the performance, we propose a supervised learning strategy. Specifically, we set up a structured learning framework for conducting supervised personalized diversification, in which we add features extracted directly from tokens of documents and those utilized by unsupervised personalized diversification algorithms, and, importantly, those generated from our proposed user-interest latent Dirichlet topic model. We also define two constraints in our structured learning framework to ensure that search results are both diversified and consistent with a user's interest. To further boost the efficiency of training, we propose a fast training framework for our proposed method by adding additional multiple highly violated but also diversified constraints at every training iteration of the cutting-plane algorithm. We conduct experiments on an open dataset and find that our supervised learning strategy outperforms unsupervised personalized diversification methods as well as other plain personalization and plain diversification methods. Our fast training framework significantly saves training time while it maintains almost the same performance.

Query auto-completion (QAC) is a prominent feature of modern search engines. It is aimed at saving user's time and enhancing the search experience. Current QAC models mostly rank matching QAC candidates according to their past popularity, i.e., frequency. However, query popularity changes over time and may vary drastically across users. Hence, rankings of QAC candidates should be adjusted accordingly. In previous work time-sensitive QAC models and user-specific QAC models have been developed separately. Both types of QAC model lead to important improvements over models that are neither time-sensitive nor personalized. We propose a hybrid QAC model that considers both of these aspects: time-sensitivity and personalization.

We tackle the problem of searching microblog posts and frame it as a rank aggregation problem where we merge result lists generated by separate rankers so as to produce a final ranking to be returned to the user. We propose a rank aggregation method, TimeRA, that is able to infer the rank scores of documents via latent factor modeling. It is time-aware and rewards posts that are published in or near a burst of posts that are ranked highly in many of the lists being aggregated. Our experimental results show that it significantly outperforms state-of-the-art rank aggregation and time-sensitive microblog search algorithms.

This paper is concerned with the problem of personalized diversification of search results, with the goal of enhancing the performance of both plain diversification and plain personalization algorithms. In previous work, the problem has mainly been tackled by means of unsupervised learning. To further enhance the performance, we propose a supervised learning strategy. Specifically, we set up a structured learning framework for conducting supervised personalized diversification, in which we add features extracted directly from the tokens of documents and those utilized by unsupervised personalized diversification algorithms, and, importantly, those generated from our proposed user-interest latent Dirichlet topic model. Based on our proposed topic model whether a document can cater to a user's interest can be estimated in our learning strategy. We also define two constraints in our structured learning framework to ensure that search results are both diversified and consistent with a user's interest. We conduct experiments on an open personalized diversification dataset and find that our supervised learning strategy outperforms unsupervised personalized diversification methods as well as other plain personalization and plain diversification methods.

A query considered in isolation provides limited information about the searcher's interest. Previous work has considered various types of user behavior, e.g., clicks and dwell time, to obtain a better understanding of the user's intent. We consider the searcher's search and page view history. Using search logs from a commercial search engine, we (i) investigate the impact of features derived from user behavior on reranking a generic ranked list; (ii) optimally integrate the contributions of user behavior and candidate documents by learning their relative importance per query based on similar users. We use dwell time on clicked URLs when estimating the relevance of documents for a query, and perform Bayesian Probabilistic Matrix Factorization as smoothing to predict the relevance. Considering user behavior achieves better rankings than non-personalized rankings. Aggregation of user behavior and query-document features with a user-dependent adaptive weight outperforms combinations with a fixed uniform value.

A popular strategy for search result diversification is to first retrieve a set of documents utilizing a standard retrieval method and then rerank the results. We adopt a different perspective on the problem, based on data fusion. Starting from the hypothesis that data fusion can improve performance in terms of diversity metrics, we examine the impact of standard data fusion methods on result diversification. We take the output of a set of rankers, optimized for diversity or not, and find that data fusion can significantly improve state-of-the art diversification methods. We also introduce a new data fusion method, called diversified data fusion, which infers latent topics of a query using topic modeling, without leveraging outside information. Our experiments show that data fusion methods can enhance the performance of diversification and DDF significantly outperforms existing data fusion methods in terms of diversity metrics.

Hierarchical multi-label classification assigns a document to multiple hierarchical classes. In this paper we focus on hierarchical multi-label classification of social text streams. Concept drift, complicated relations among classes, and the limited length of documents in social text streams make this a challenging problem. Our approach includes three core ingredients: short document expansion, time-aware topic tracking, and chunk-based structural learning. We extend each short document in social text streams to a more comprehensive representation via state-of-the-art entity linking and sentence ranking strategies. From documents extended in this manner, we infer dynamic probabilistic distributions over topics by dividing topics into dynamic "global" topics and "local" topics. For the third and final phase we propose a chunk-based structural optimization strategy to classify each document into multiple classes. Extensive experiments conducted on a large real-world dataset show the effectiveness of our proposed method for hierarchical multi-label classification of social text streams.

The task of finding groups is a natural extension of search tasks aimed at retrieving individual entities. We introduce a group finding task: given a query topic, find knowledgeable groups that have expertise on that topic. We present four general strategies to this task. The models are formalized using generative language models. Two of the models aggregate expertise scores of the experts in the same group for the task, one locates documents associated with experts in the group and then determines how closely the documents are associated with the topic, whilst the remaining model directly estimates the degree to which a group is a knowledgeable group for a given topic. We construct a test collections based on the TREC 2005 and 2006 Enterprise collections. We find significant differences between different ways of estimating the association between a topic and a group. Experiments show that our knowledgeable group finding models achieve high absolute scores.

We focus on the problem of selecting meaningful tweets given a user's interests; the dynamic nature of user interests, the sheer volume, and the sparseness of individual messages make this an challenging problem. Specifically, we consider the task of time-aware tweets summarization, based on a user's history and collaborative social influences from ``social circles.'' We propose a time-aware user behavior model, the Tweet Propagation Model (TPM), in which we infer dynamic probabilistic distributions over interests and topics. We then explicitly consider novelty, coverage, and diversity to arrive at an iterative optimization algorithm for selecting tweets. Experimental results validate the effectiveness of our personalized time-aware tweets summarization method based on TPM.